* [Caml-list] Overhead of first class modules
@ 2011-09-16 20:08 Alex Rubinsteyn
2011-09-16 20:36 ` ygrek
0 siblings, 1 reply; 2+ messages in thread
From: Alex Rubinsteyn @ 2011-09-16 20:08 UTC (permalink / raw)
To: caml-list
[-- Attachment #1: Type: text/plain, Size: 972 bytes --]
I ran a very simple experiment gauging the overhead of OCaml's new
first-class modules.
*direct.ml: *
let f x = x + 1
let _ =
let acc = ref 0 in
for i = 0 to 10000000 do
acc := !acc + (f 1)
done
*first_class_module.ml*
module type S = sig val f : int -> int end
module M = struct let f x = x + 1 end
let package = (module M : S)
let _ =
let acc = ref 0 in
for i = 0 to 10000000 do
let module Local = (val package : S) in
acc := !acc + (Local.f 1)
done
obj.ml
let o = object method f x = x + 1 end
let _ =
let acc = ref 0 in
for i = 0 to 10000000 do
acc := !acc + o#f 1
done
# ocamlopt direct.ml -o direct
# ocamlopt first_class_module.ml -o first_class_module
# ocamlopt obj.ml -o obj
# time ./direct
real 0m0.008s
user 0m0.010s
# time ./first_class_module
real 0m0.012s
user 0m0.020s
# time ./obj
real 0m0.058s
user 0m0.060s
So, the overhead of first class modules is slight, especially when compared
with method invocation.
[-- Attachment #2: Type: text/html, Size: 2436 bytes --]
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [Caml-list] Overhead of first class modules
2011-09-16 20:08 [Caml-list] Overhead of first class modules Alex Rubinsteyn
@ 2011-09-16 20:36 ` ygrek
0 siblings, 0 replies; 2+ messages in thread
From: ygrek @ 2011-09-16 20:36 UTC (permalink / raw)
To: caml-list
On Fri, 16 Sep 2011 16:08:55 -0400
Alex Rubinsteyn <alex.rubinsteyn@gmail.com> wrote:
> I ran a very simple experiment gauging the overhead of OCaml's new
> first-class modules.
This is not a valid experiment. In first two cases ocamlopt inlines
function call. Also mind that in this code ocamlopt calls the
Local.f function directly, not through the module.
But let's correctly microbenchmark the cost of direct call vs
indirect vs message sending :)
let n = 100_000_000
let f x = x + 1
let direct () =
let acc = ref 0 in
for i = 0 to n do
acc := !acc + (f 1)
done
module type S = sig val f : int -> int end
module M = struct let f x = x + 1 end
let package = ref (module M : S)
let fstcls () =
let acc = ref 0 in
for i = 0 to n do
let module Local = (val !package : S) in
acc := !acc + (Local.f 1)
done
let o = object method f x = x + 1 end
let obj () =
let acc = ref 0 in
for i = 0 to n do
acc := !acc + o#f 1
done
let () =
match Sys.argv with
| [|_;"direct"|] -> direct ()
| [|_;"fstcls"|] -> fstcls ()
| [|_;"obj"|] -> obj ();
| _ -> failwith "ORLY?"
Disable inlining :
/opt/ocaml-3.12.1/bin/ocamlopt -inline 0 -S speed.ml -o speed
Run (Linux 3.0.0-1-686-pae i686) :
$ time ./speed direct
real 0m0.711s
user 0m0.560s
sys 0m0.012s
$ time ./speed fstcls
real 0m0.631s
user 0m0.508s
sys 0m0.000s
$ time ./speed obj
real 0m1.157s
user 0m0.840s
sys 0m0.016s
Looks like it is time to sprinkle some performance-critical code with first class modules :)
--
ygrek
http://ygrek.org.ua
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2011-09-16 20:37 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-09-16 20:08 [Caml-list] Overhead of first class modules Alex Rubinsteyn
2011-09-16 20:36 ` ygrek
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).